Skip to content

Writing classifier article

Here you'll find instructions of how to create an article that describes the classifier you developed, including the upload of the source code itself. As an example, the Weka classifier ZeroR is used.

Preparation#

The content of the wiki is available as repository on GitHub, which also contains a guide on how to build and test the repository using MkDocs. Add or modify items as necessary and then do a pull request. See this link for details on writing articles.

Article Layout#

To provide fast access to a certain classifier, the following topics should be covered in an article about a classifier:

  • insert a full description of your classifier, whether it handles numerical and/or nominal classes/attributes; in a nutshell, information a potential user needs to decide whether this specific classifier is suitable for a certain problem
     # Description
     Class for building and using a 0-R classifier. Predicts the mean
     (for a numeric class) or the mode (for a nominal class).
    
  • add the reference paper, if any
     # Reference
     -none-
    
  • add the package the classifier belongs to, that people can adapt their GenericObjectEditor props file and use the classifier within Weka.
     # Package
     weka.classifiers.rules
    
  • add the link to download the source code, either an internal one, if you uploaded the source code, or an external one, if you want to point to a different resource; here it is an internal one, pointing to ZeroR.java
     # Download
     Source code: [ZeroR.java](files/ZeroR.java)
    
  • (optional) add additional information, e.g., some benchmarks; which is nothing in our case
     # Additional Information
     -none-
    
  • finally, save and deploy the article